diff --git a/tests/unit/services/test_cluster_state_metric_collectors.py b/tests/unit/services/test_cluster_state_metric_collectors.py
new file mode 100644
index 0000000..cd1a249
--- /dev/null
+++ b/tests/unit/services/test_cluster_state_metric_collectors.py
@@ -0,0 +1,118 @@
+from __future__ import annotations
+
+from ariadne.services import cluster_state_metric_collectors as collectors
+
+
+def test_collect_vm_core_records_query_errors(monkeypatch) -> None:
+    def boom(_expr):
+        raise RuntimeError("vm unavailable")
+
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_vm_scalar", boom)
+
+    collectors._collect_vm_core(metrics, errors)
+
+    assert errors == ["vm: vm unavailable"]
+
+
+def test_collect_node_metrics_builds_and_reports_baselines(monkeypatch) -> None:
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_postgres_connections", lambda errors: {"used": 2})
+    monkeypatch.setattr(collectors, "_hottest_nodes", lambda errors: {"cpu": {"node": "titan-1"}})
+    monkeypatch.setattr(
+        collectors,
+        "_node_usage",
+        lambda errors: {"cpu": [{"value": 2.0}], "ram": [], "net": [], "io": [], "disk": []},
+    )
+    monkeypatch.setattr(collectors, "_usage_stats", lambda entries: {"count": len(entries)})
+    monkeypatch.setattr(collectors, "_node_usage_exprs", lambda: {"cpu": "expr"})
+    monkeypatch.setattr(collectors, "_vm_baseline_map", lambda _expr, _label, _window: {"titan-1": {"avg": 2.0}})
+    monkeypatch.setattr(
+        collectors,
+        "_baseline_map_to_list",
+        lambda baseline, label: [{label: name, **stats} for name, stats in baseline.items()],
+    )
+
+    collectors._collect_node_metrics(metrics, errors)
+
+    assert metrics["node_usage_stats"]["cpu"] == {"count": 1}
+    assert metrics["node_baseline"]["cpu"] == [{"node": "titan-1", "avg": 2.0}]
+    assert metrics["node_baseline_map"]["titan-1"]["cpu"] == {"avg": 2.0}
+    assert errors == []
+
+
+def test_collect_node_metrics_records_baseline_errors(monkeypatch) -> None:
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_postgres_connections", lambda errors: {})
+    monkeypatch.setattr(collectors, "_hottest_nodes", lambda errors: {})
+    monkeypatch.setattr(collectors, "_node_usage", lambda errors: {})
+    monkeypatch.setattr(collectors, "_usage_stats", lambda entries: {})
+    monkeypatch.setattr(collectors, "_node_usage_exprs", lambda: (_ for _ in ()).throw(RuntimeError("baseline down")))
+
+    collectors._collect_node_metrics(metrics, errors)
+
+    assert errors == ["baseline: baseline down"]
+
+
+def test_collect_trend_issue_and_alert_errors(monkeypatch) -> None:
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_build_metric_trends", lambda *args: (_ for _ in ()).throw(RuntimeError("trend down")))
+    monkeypatch.setattr(collectors, "_namespace_reason_entries", lambda *args: (_ for _ in ()).throw(RuntimeError("issue down")))
+    monkeypatch.setattr(collectors, "_vm_alerts_now", lambda: (_ for _ in ()).throw(RuntimeError("alerts down")))
+
+    collectors._collect_trend_metrics(metrics, errors)
+    collectors._collect_issue_metrics(metrics, errors)
+    collectors._collect_alert_metrics(metrics, errors)
+
+    assert "trends: trend down" in errors
+    assert "issues: issue down" in errors
+    assert "alerts: alerts down" in errors
+
+
+def test_collect_namespace_metrics_builds_baselines_and_totals(monkeypatch) -> None:
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_vm_vector", lambda _expr: [])
+    monkeypatch.setattr(collectors, "_filter_namespace_vector", lambda entries: entries)
+    monkeypatch.setattr(collectors, "_vm_namespace_totals", lambda expr: {"apps": 2.0} if "cpu" in expr else {"apps": 4.0})
+    monkeypatch.setattr(collectors, "_build_namespace_capacity", lambda *args: [{"namespace": "apps", "cpu_usage_ratio": 2.0}])
+    monkeypatch.setattr(collectors, "_namespace_usage_exprs", lambda: {"cpu": "expr"})
+    monkeypatch.setattr(collectors, "_vm_baseline_map", lambda _expr, _label, _window: {"apps": {"avg": 2.0}})
+    monkeypatch.setattr(
+        collectors,
+        "_baseline_map_to_list",
+        lambda baseline, label: [{label: name, **stats} for name, stats in baseline.items()],
+    )
+    monkeypatch.setattr(collectors, "_namespace_capacity_summary", lambda capacity: {"count": len(capacity)})
+
+    collectors._collect_namespace_metrics(metrics, errors)
+
+    assert metrics["namespace_totals"]["cpu"] == [{"namespace": "apps", "value": 2.0}]
+    assert metrics["namespace_baseline"]["cpu"] == [{"namespace": "apps", "avg": 2.0}]
+    assert metrics["namespace_baseline_map"]["apps"]["cpu"] == {"avg": 2.0}
+    assert metrics["namespace_capacity_summary"] == {"count": 1}
+    assert errors == []
+
+
+def test_collect_namespace_metrics_records_usage_and_baseline_errors(monkeypatch) -> None:
+    metrics: dict[str, object] = {}
+    errors: list[str] = []
+    monkeypatch.setattr(collectors, "_vm_vector", lambda _expr: (_ for _ in ()).throw(RuntimeError("usage down")))
+    monkeypatch.setattr(collectors, "_namespace_usage_exprs", lambda: (_ for _ in ()).throw(RuntimeError("baseline down")))
+    monkeypatch.setattr(collectors, "_namespace_capacity_summary", lambda capacity: {"count": len(capacity)})
+
+    collectors._collect_namespace_metrics(metrics, errors)
+
+    assert "namespace_usage: usage down" in errors
+    assert "baseline: baseline down" in errors
+    assert metrics["namespace_capacity_summary"] == {"count": 0}
+
+
+def test_namespace_totals_list_filters_invalid_names() -> None:
+    assert collectors._namespace_totals_list({"apps": 2.0, "": 9.0, 3: 4.0}) == [
+        {"namespace": "apps", "value": 2.0}
+    ]